In a 8-week production-scale experiment at a commercial trout farm, the effects of dietary lipid level and phosphorus level on phosphorus (P) and nitrogen (N) utilization of rainbow trout (initial mean weight 99 g) were assessed. A low-phosphorus, high-lipid experimental diet (457 g protein, 315 g lipid, 9.1 g P kg–1 dry diet) was compared with a commonly used commercial diet (484 g protein, 173 g lipid, 13.6 g P kg–1 dry diet). P and N budgets were constructed using data from the production-scale experiment and digestibility data for the two diets. In addition, orthophosphate and ammonia-N waste were measured in effluent over one 24-h period. Relative to the commercial diet, the experimental diet resulted in significantly increased feed efficiency ratio, N retention and P retention, and substantially reduced dissolved, solid and total P waste (g kg–1 dry feed). Although N retention resulting from the experimental diet was higher, this was attributable to higher N (protein) digestibility of the experimental diet. Solid N waste (g kg–1 dry feed) resulting from the experimental diet was substantially lower, but dissolved N waste (g kg–1 dry feed) was not significantly different relative to the commercial diet. Mean effluent orthophosphate production (mg day–1 kg–1 fish) of fish fed the experimental diet was substantially lower than that of fish fed the commercial diet (P < 0.05), but effluent ammonia-N production (mg day–1 kg–1 fish) was not significantly affected by dietary treatment.